Customer demand usually fluctuates in a random fashion. Therefore, it is modeled in this simulation as normally distributed random noise with a mean value of mean=1000 and a standard deviation of stdev=100. The noise is sampled once per week and kept constant for the corresponding week. The order flow is modeled using the equation:

RRR(t) = RRRini + normal(1000,100);

Simulate the model across 10 years (520 weeks), and plot on a single graph the incoming orders, the production flow in the factory, and the levels of goods in retail, distribution, and the factory as functions of time:

Choose Radau-IIa as your integration algorithm. It handles noise input better than DASSL.